2 Feb
2023
2 Feb
'23
8:11 a.m.
On Feb 2, 2023, at 11:03 AM, Sellam Abraham via cctalk
<cctalk(a)classiccmp.org> wrote:
This discussion on core memory got me thinking:
Is it possible to "read" core memory by examining each core using some kind
of instrument that would sense its "charge" (or lack thereof)
non-destructively?
Could a piece of paper be placed over a core plane and fine particles of
iron sprinkled onto the paper to literally "see" the bits?
Please excuse me for any ignorant assumptions made here on my part.
Sellam
That's a very good question. The difficulty with doing this isn't that the core
states are "magnetized" and "not magnetized" but rather they are
"magnetized in one direction" and "magnetized in the other direction".
So flux-visualizing won't help because 1 and 0 will look the same.
The other problem is that a toroidal core, by design, constrains nearly all the magnetic
flux inside the core material, with very little "leakage".
That said, it would be an interesting experiment to take a core memory module and hold a
small Hall effect probe up against the cores, especially inside the core opening. I
don't know if those sensors can be sensitive enough to pick up the leakage flux. If
yes, it may well depend on the core diameter; the later generations tended to have smaller
cores for both density and speed. CDC mainframe cores are probably small for the time, I
don't remember seeing the numbers.
paul